Purchase this article with an account.
S. Bearelly, M. Z. Chen, M. P. Kelly, S. W. Cousins; Comparing Fundus Autofluorescence Imaging With Fundus Photographs in Patients With Geographic Atrophy Related to Age-Related Macular Degeneration. Invest. Ophthalmol. Vis. Sci. 2007;48(13):2163. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Prior studies of geographic atrophy (GA) have used color fundus photos to measure size and progression of atrophy. Fundus autofluorescence (FAF) imaging of absent autofluorescence is another method for determination of GA area. We sought to compare the consistency between the two methods in terms of measurement of the area of GA.
We performed a retrospective review of 30 eyes (21 patients) with confirmed diagnosis of age-related macular degeneration with geographic atrophy. All patients had color fundus photographs (Visupac, Carl Zeiss Meditec) and FAF images recorded with the Heidelberg Retina Angiograph confocal scanning laser ophthalmoscope (Heidelberg Engingeering, Germany). Color images were converted to grayscale and registered with FAF images using ImageJ software. Registered images were analyzed by two graders who compared and graded the extent of GA evident on FAF images and fundus photos. Photos were saved without any private health information, and this study met our Institutional Review Board’s exemption criteria.
The average age was 78.7 years (range 65-89). 12 patients (57%) were female. Only 5 of 30 eyes (17%) were found to have equivalent GA determined by both methods. 14 FAF images (47%) were found to have more GA as compared with fundus photos and 9 (30%) were found to have less GA. Two of 30 images (7%) were found to have more GA in some areas and less in others. In general, FAF image grading was more reproducible and consistent between graders. Confounding findings on photographic images that caused disagreement with FAF grading included the presence of extensive regressed drusen (under-estimated FAF-determined GA) or depigmentation of the retinal pigment epithelium (over-estimated FAF-determined GA).
Our results indicate a significant disagreement between fundus photos and FAF in determination of the extent of GA. Reasons for inconsistencies between the two methods include difficulty in photographic interpretation of depigmentation (which results in over-estimation of areas of GA), difficulty in identifying GA in zones of confluent regressed drusen (resulting in undercalling of GA), and failure of photographs to identify small (< 250 micron) satellite areas of GA. FAF imaging with the Heidelberg Retinal Angiograph seems to be a consistent and reproducible method for the measurement of GA.
This PDF is available to Subscribers Only